Dr Dev Kambhampati | New Insights on Retail Ecommerce

New Insights on Retail E-Commerce
Office of the Chief Economist  Economics and Statistics Administration  Page 1
Executive Summary
The U.S. Census Bureau has been collecting data on retail sales since the
1950s and data on e-commerce retail sales since 1998. As the Internet
has become ubiquitous, many retailers have created websites and even
entire divisions devoted to fulfilling online orders. Many consumers have
turned to e-commerce as a matter of convenience or to increase the
variety of goods available to them. Whatever the reason, retail e-
commerce sales have skyrocketed and the Internet will undoubtedly
continue to influence how consumers shop, underscoring the need for
good data to track this increasingly important economic activity.
In June 2017, the Census Bureau released a new supplemental data table
on retail e-commerce by type of retailer. The Census Bureau developed
these estimates by re-categorizing e-commerce sales data from its
existing “electronic shopping” sales data according to the primary
business type of the retailer, such as clothing stores, food stores, or
electronics stores. This report examines how the new estimates enhance
our understanding of where consumers are shopping online and also
provides an overview of trends in retail and e-commerce sales. Findings
from this report include:
 E-commerce sales accounted for 7.2 percent of all retail sales in 2015,
up dramatically from 0.2 percent in 1998.
 E-commerce sales have been growing nine times faster than traditional
in-store sales since 1998.
 In 2015, 87 percent of total retail e-commerce sales, or $294.8 billion,
were attributed to electronic shopping and mail order houses, which
includes both Internet-only businesses and traditional stores’ online
divisions. The new data released by Census shows that nonstore
retailers, or businesses with little to no physical store establishments,
accounted for 65 percent, or $192.1 billion, of these sales, while the
online and catalog divisions of traditional retailers accounted for the
remaining $102.7 billion.
U.S. Department of Commerce
Economics and Statistics Administration
Office of the Chief Economist
By
Jessica R. Nicholson
ESA Issue Brief
#04-17
July 26, 2017

 When e-commerce sales from electronic shopping and mail order houses are attributed to the
primary business activity of their brick-and-mortar stores, e-commerce sales accounted for 18
percent of total sales from electronics and appliance stores, 10 percent from miscellaneous
store retailers, and 10 percent from clothing and clothing accessories stores.
 Census data are also available for e-commerce sales by type of product, and these data show
that consumers purchase a wide range of products online. In 2015, “other merchandise,”
including collectibles, souvenirs, auto parts and accessories, hardware, lawn and garden
equipment and supplies, and jewelry, accounted for the largest share of e-commerce sales from
online businesses at $52.9 billion, or 18 percent of all retail e-commerce from these types of
businesses. Clothing and clothing accessories accounted for another 18 percent at $52.1 billion.
Online consumers also purchased $28.9 billion of furniture and home furnishings (10 percent of
the total) and $26.1 billion of electronics and appliances (9 percent).

Introduction
The U.S. Census Bureau has been collecting data on retail sales since the 1950s and data on e-commerce
retail sales since 1998. As the Internet has become ubiquitous, many retailers have created websites and
even entire divisions devoted to fulfilling online orders. Consumers have turned to e-commerce as a
matter of convenience or to increase the variety of goods available to them. Whatever the reason, retail
e-commerce growth is skyrocketing and the Internet will undoubtedly continue to impact how
consumers shop. The growing importance of online shopping drives the need for good data on this
economic activity.
In June 2017, the Census Bureau released new data that provides more detail on the types of retailers
that consumers are frequenting online. The Census Bureau developed the new estimates by re-
categorizing e-commerce sales from its “electronic shopping” category according to the primary
business type of the retailer, such as clothing stores, food stores, or electronics stores. By presenting the
data in this new way, retailers, policymakers, and the public can better identify the types of stores
where consumers are spending their retail dollars online. This report looks at trends in retail and e-
commerce sales and looks at how the new estimates enhance our understanding of online shopping.
Overview of retail sales and retail e-commerce data
The retail sector includes businesses classified in sectors 44 and 45 of the North American Industrial
Classification System (NAICS)1
(See Table 1). The Census Bureau collects sales data on this sector using
three different surveys: the Advance Monthly Retail Trade Survey (MARTS); the Monthly Retail Trade
Survey (MRTS); and the Annual Retail Trade Survey (ARTS)2
and publishes this data by type of business.3
To be included in the surveys, companies must have at least one establishment and sell merchandise to
final consumers.4
The Census Bureau defines “e-commerce” as the sale of goods and services where the buyer places an
order or the price and terms of the sale are negotiated over the Internet, a mobile device (M-
commerce), extranet, Electronic Data Interchange, electronic mail, or another comparable online
system. Payment may or may not be made online.5
E-commerce is not a type of business classified by
1
For more information on NAICS 44-45, see the Census Bureau Website at: https://www.census.gov/cgi-
bin/sssd/naics/naicsrch?chart_code=44&search=2017 NAICS Search.
2
For more information on these surveys, see the Census Bureau Website at:
https://www.census.gov/retail/index.html. Every five years, the Census Bureau conducts an Economic Census of
Retail Trade which is used to benchmark the annual data. For current survey forms, see the Census Bureau
Website at: https://www.census.gov/retail/mrts/get_forms.html.
3
Data on retail sales are available down to the 6-digit NAICS code level for some industries and at a less detailed
level for others.
4
https://www.census.gov/retail/marts/about_the_surveys.html
5
General definition that covers e-commerce collection at the Census Bureau, available at:
https://www.census.gov/retail/definitions.html. The Census Bureau also publishes data on e-commerce in
manufacturing, merchant wholesale trade, and selected service industries in their annual e-stats report. The latest
report with 2015 data is available at:
https://www.census.gov/content/dam/Census/library/publications/2017/econ/e15-estats.pdf.

NAICS, but rather a means by which customers can purchase goods. Conceivably, with perhaps the
exception of gasoline stations, all of the types of businesses listed in Table 1 could sell goods online.
In 1998, the Census Bureau began collecting and publishing quarterly data on retail e-commerce sales.6
The survey forms used to collect this data ask businesses for the value of their e-commerce sales in
dollars or as a percentage of total retail sales. Total annual retail sales and e-commerce sales from 1998
through 2015 are shown in Figure 1. In 2015, annual retail sales in the United States totaled $4.7 trillion.
That same year, retail e-commerce sales were estimated to be $340.4 billion and accounted for 7.2
percent of total retail sales. Although e-commerce sales are a relatively small portion of total retail sales,
they have been an important source of sales growth for retailers. In 1998, retail e-commerce sales
totaled only $5.0 billion, or 0.2 percent of the $2.6 trillion in retail sales that year. From 1998 to 2015,
total retail sales grew by $2.1 trillion—and 16 percent of that increase, or $335.4 billion, was e-
commerce.
6
E-commerce data are collected from the same sample used for the MRTS. Advance U.S. retail sales are estimated
from a subsample of the MRTS sample that is not of adequate size to measure changes in retail e-commerce sales.
See Quarterly E-Commerce Methodology for more information, available at:
https://www.census.gov/retail/ecommerce/how_surveys_are_collected.html.
Sector
Retail trade
Industries
Motor vehicle and parts dealers
Furniture and home furnishing stores
Electronics and appliance stores
Building material and garden equipment supplies dealers
Food and beverage stores
Health and personal care stores
Gasoline stations
Clothing and clothing accessories stores
Sporting goods, hobby, musical instrument, and book stores
General merchandise stores
Miscellaneous store retailers
Nonstore retailers
4541 Electronic shopping and mail order houses
453
454
Table 1. Industries included in the retail trade sector
Source: Census Bureau
445
446
447
448
451
452
NAICS Code
44-45
441
442
443
444

Indexing retail e-commerce sales and all other (or non-e-commerce) retail sales to 1998 (1998=100)
clearly shows the exceptional growth of e-commerce compared to traditional brick-and-mortar retail
(See Figure 2). The index for non-e-commerce retail sales rose from 100 in 1998 to 170 in 2015, an
average of just 3 percent per year. The index for e-commerce sales increased from 100 in 1998 to a
whopping 6830 in 2015—an annual growth rate of 28 percent, nine times the growth of all other retail
sales.

Retail sales and retail e-commerce sales by type of business
As more consumers move to purchasing more goods online, knowing whether retail purchases are in
traditional stores or online will become essential to understanding the U.S. consumer marketplace.
The new Census Bureau data on e-commerce sales, discussed in more detail below, provides us with a
better picture of these sales across the various retail sales categories.
Figure 3 displays previously available annual retail sales data for 2011 through 2015, broken down by
retail e-commerce (in orange) and non-e-commerce sales (in blue) based on the data that had been
previously available. Each year, consumers spent the most money, around a fifth of all retail spending, at
motor vehicles and parts dealers. Food and beverage stores and general merchandise stores each
accounted for around 15 percent of retail sales. From 2011 to 2014, gasoline stations ranked fourth, but
with recent declines in gasoline prices, sales at gasoline stations fell by 17.6 percent in 2015, pushing
this category into the number 5 spot.

Nonstore retailers (NAICS 454) are establishments such as mail-order houses, vending machine
operators, home delivery sales, door-to-door sales, party plan sales, electronic shopping, and sales
through portable stalls (e.g., street vendors, except food).7
In 2015, nonstore retailers’ sales totaled
$509.3 billion, of which $433.7 billion came from a subgroup of this industry, electronic shopping and
mail order houses (NAICS 4541). The electronic shopping and mail order houses subgroup includes
businesses without traditional brick-and-mortar locations, but it also includes businesses which have
developed a separate online division within the company and do not fulfill online orders directly from
physical store locations—an exception to this is when stores offer online ordering with in-store or
curbside pick-up; because customers order online, the sales would be classified under electronic
shopping and mail order houses.8
For example, if a clothing store fulfills online orders from a warehouse
and also sells clothes at the local shopping mall, its e-commerce orders would be classified under
electronic shopping and mail order houses (NAICS 4541) while its mall store sales would be classified
under the primary business activity, which is clothing and clothing accessory stores (NAICS 448). This is
similar to how companies would split reporting between two distinct brick-and-mortar divisions, such as
a company that owns both grocery stores and gasoline stations, for example.
Because most retailers have in fact developed online divisions within their businesses, virtually all e-
commerce sales, as shown in Figure 3, are counted in this NAICS category rather than the primary
business activity of their brick-and-mortar stores.9
In fact, because of the way Census classifies a
businesses’ e-commerce sales, it appears that most of the retail segments displayed in Figure 3 have
virtually no e-commerce sales at all. Therefore, for all business types except nonstore retailers, e-
commerce sales appeared to account for 3 percent or less of retail sales each year from 2011 through
2015.
The new Census Bureau estimates released in June provide a better picture of actual e-commerce sales
across the various retail sales categories. The new estimates attribute the e-commerce sales of
traditional retailers with online divisions to the primary type of business of the parent company rather
than as sales by electronic shopping and mail order houses (See Figure 4).10
This allows data users to see
what portion of electronic shopping sales are from the various types of businesses with brick-and-
mortar establishments and separate online divisions, versus from mail order houses and companies that
sell online but don’t have brick-and-mortar presences.11
7
For the entire description of NAICS 454, see: https://www.census.gov/cgi-
bin/sssd/naics/naicsrch?code=454&search=2017 NAICS Search.
8
Information taken from MRTS FAQs, available at: https://www.census.gov/retail/mrts_general_faqs.html.
9
In 2015, out of total e-commerce sales of $340.4 billion, $294.8 billion, or 86.6 percent were in NAICS 4541.
10
Supplemental U.S. Sales—Total and E-commerce Sales by Primary Business Activity—2011-2015, available at:
http://www2.census.gov/retail/releases/current/arts/supecommerce4541.xlsl.
11
For companies with separate store and e-commerce components as described above, the new supplemental e-
commerce table reallocates the sales of the NAICS 4541 component to the primary 3-digit NAICS code of the brick-
and-mortar component of the company. Companies without a brick-and-mortar component remain classified
under NAICS 454. For more information about the new supplemental table and data limitations, see Latest Annual
Retail Trade Report, Supplemental U.S. Sales—Total and E-commerce Sales by Primary Business Activity—2011-
2015, Explanatory Material, available at:
http://www2.census.gov/retail/releases/current/arts/sup_ec_tech_doc.doc.

In 2015, $294.8 billion, or 68 percent of all retail sales from electronic shopping and mail order houses,
were e-commerce sales, up from 57 percent in 2011. Because of the new estimates, we now know that,
as shown in Figure 4, $192.1 billion of these e-commerce sales in 2015 can be attributed to retailers that
almost exclusively sell items online or through means other than traditional stores.12
Consumers
purchased the remaining $102.7 billion online but from all of the other business types presented in
Figure 4 that also had brick-and-mortar stores. In addition, for all of the types of retailers shown in
Figure 4, with the exception of nonstore retailers and health and personal care stores, about 90 percent
or more of sales through electronic shopping and mail order houses were e-commerce sales in 2015.13
12
The survey sample was not designed to produce NAICS 4541 estimates by primary business activity. Therefore,
the amount of sampling error and nonsampling error with some estimates may be substantial. In addition, the
variance estimation method has limitations when being applied to the supplemental e-commerce estimates. As a
result, the produced measures of reliability in the supplemental table and Figure 4 may underestimate the
sampling variability. The U.S. Census Bureau recommends that individuals use these estimates with caution, as
sampling error and nonsampling error could affect the conclusions drawn from the estimates.
13
The value of non-e-commerce sales from electronic shopping and mail order houses attributed to health and
personal care stores are likely mail order prescription services.

Figure 5 incorporates the data from electronic shopping and mail order houses that are allocated to
retailers with traditional brick and mortar stores back into the overall retail sales data from Figure 3.
Although the majority of retail sales from the electronic shopping and mail order houses category
remains within nonstore retailers, the reallocation of the remaining sales boost, to some degree, the e-
commerce share of sales in other types of businesses, as can be seen by the now-visible orange bars for
the other business types in Figure 5.
Recall that before this reallocation, e-commerce sales accounted for 3 percent or less of retail sales in
every type of business, except for nonstore retailers, each year from 2011 through 2015. However, after
the reallocation, e-commerce sales accounted for 14 percent of retail sales from electronics and
appliance stores in 2011. This share increased to 18 percent by 2015. In clothing and clothing
accessories stores, the share of e-commerce sales after the reallocation was 7 percent in 2011 and up to
10 percent in 2015. Other retailers who had significant e-commerce sales were: miscellaneous store
retailers (10 percent e-commerce sales in 2015); sporting goods, hobby, book and music stores (8
percent); and furniture and home furnishings stores (6 percent). The reallocation of the retail sales data

from electronic shopping and mail order houses to other business types shows that e-commerce is an
important, and growing, method for consumer purchases for many types of retailers, not just nonstore
retailers.
Retail e-commerce sales by merchandise line
While Figures 4 and 5 give us a better sense of the extent to which various types of stores are selling
online, they only provide limited information about the types of products that are purchased online. It is
safe to assume that online sales from clothing and clothing accessories stores mainly involve clothing
items, but there is no safe assumption about the type of products involved in the online sales of
electronic shopping businesses (included in NAICS 4541) that sell almost exclusively over the Internet.14
As an additional resource on retail e-commerce, the Census Bureau, since 1999, has been collecting and
14
Data on retail sales by merchandise line are only available for electronic shopping and mail order houses.
Therefore, while the merchandise line data provide insight to what consumers are buying online, the majority of
the retail sales data published by Census are by indicator business type, making the merchandise line data not
directly comparable.

publishing data on total sales and e-commerce sales by merchandise line, or type of good, for electronic
shopping and mail order houses.15
Figure 6 displays the e-commerce sales of electronic shopping and
mail order houses by the type of product consumers purchase online, as opposed to the types of
retailer selling online.
Purchases of clothing and clothing accessories (including footwear) and “other merchandise,” each with
18 percent of the total, are the categories accounting for the largest shares of total online purchases
from electronic shopping and mail order houses. While “other merchandise” includes a variety of items
such as jewlery, lawn equipment, and collectibles, the clothing and clothing accessories category
provides more insights about online purchases. In 2015, online consumers purchased $52.1 billion of
clothing and clothing accessories online from electronic shopping and mail order houses.16
These
shoppers purchased another $4.1 billion in merchandise from clothing and clothing accessories retailers
(shown in Figure 3, but difficult to see because of the relatively small value).
15
U.S. Electronic Shopping and Mail-Order Houses (NAICS 4541) - Total and E-commerce Sales by Merchandise Line
(1999-2015) available at: http://www2.census.gov/retail/releases/current/arts/ecommerce4541.xls.
16
Figure 4 shows $24.2 billion in online sales reallocated from electronic shopping and mail order houses to
clothing and clothing accessories stores in 2015. This total varies greatly from the $52.1 billion in clothing and
clothing accessories purchased online as reported in the merchandise line data. This is likely partially due to the
fact that consumers can purchase clothing and clothing accessories from stores other than clothing stores. For
example, consumers can also purchase clothing and clothing accessories from general merchandise stores,
sporting goods stores, and potentially other types of retailers. Additionally, any clothing purchased from online-
only retailers would be captured in the nonstore retailer category.

Figure 7 shows the growth of e-commerce sales from electronic shopping and mail order houses by
merchandise line since 2000 (index 2000=100). All of the merchandise lines have experienced relatively
high growth. The lowest growth occurred in the sales of computer hardware, as is shown by the red
index line in Figure 7. However, even this index increased from 100 in 2000 to 289 in 2015, as e-
commerce sales of these goods increased from $4.2 billion to $17.5 billion.17
The lines at the top of the
chart (in gray, blue, and green) show the merchandise lines with the highest growth in e-commerce sales
from electronic shopping and mail order houses. The most impressive is the index for the online sales of
drugs, health, and beauty aids, which grew from 100 ($0.7 billion) in 2000 to 3063 ($21.5 billion) in
2015.
Conclusion
In the United States, most consumer purchases still take place in brick-and-mortar stores; according to
the latest Census data, e-commerce sales account for 7.2 percent of all retail sales. However, consumer
preferences for how they purchase goods are changing—and quickly. E-commerce sales are growing
rapidly and consumers are using the Internet to purchase everything from food to clothing to cars. As
these changes take place in the economy, the Census Bureau, and the Department of Commerce more
broadly, are working to improve the data available on e-commerce and other topics related to the
digital economy.
The new estimates on e-commerce by type of business is a positive step in that direction. The Census
Bureau’s ability to present existing data in a new way provides new insights on the amount of e-
commerce sales generated from online divisions of brick-and-mortar retailers so that retailers,
policymakers, and other data users can make more informed decisions. As a next step, the Census
Bureau is working to provide additional detail on sales by type of business in its quarterly retail e-
commerce sales data release.
17
Computer hardware accounted for 36 percent of retail goods sold through e-commerce in 1999. In 2015, this
merchandise line accounted for just 6 percent of retail e-commerce sales.

Acknowledgments
The author would like to thank the following persons who provided comments, suggestions, and other
contributions to this report. Any errors in the report are solely the author’s responsibility.
Economics and Statistics Administration, Office of the Chief Economist
Rob Rubinovitz, Deputy Chief Economist, Department of Commerce
Dave Langdon, Economist and Senior Policy Advisor
Ryan Noonan, Economist
Regina Powers, Economist
U.S. Census Bureau
Ian Thomas, Assistant Division Chief, Retail and Wholesale Trade Sectors
Scott Scheleur, Assistant Division Chief, Retail and Wholesale Indicators Programs
Chris Savage, Branch Chief, Retail Trade Branch

Suggested citation:
Nicholson, J. Office of the Chief Economist, Economics and Statistics Administration, U.S. Department of
Commerce. (2017). New Insights on Retail E-Commerce (ESA Issue Brief # 04-17). Retrieved from
https://www.esa.gov/reports/new-insights-retail-e-commerce.
U.S. Department of Commerce
Economics and Statistics Administration
Room 4848
1401 Constitution Ave., NW
Washington, DC 20230
www.esa.gov
Technical inquiries:
(202) 482-3523
Media inquiries:
Office of Communications
(202) 482-3331

Dr Dev Kambhampati | New Insights on Retail Ecommerce

Recommended

Recommended

More Related Content

What's hot

What's hot (17)

Similar to Dr Dev Kambhampati | New Insights on Retail Ecommerce

Similar to Dr Dev Kambhampati | New Insights on Retail Ecommerce (20)

More from Dr Dev Kambhampati

More from Dr Dev Kambhampati (20)

Recently uploaded

Recently uploaded (11)

Dr Dev Kambhampati | New Insights on Retail Ecommerce